Internet Infrastructure

Objectives
• Learn how networks are segmented to improve
performance and how hardware devices work to
segment a network
• Investigate the way large networks are divided logically
into subnets to make them easier to manage and
improve performance
• Study data routing across the Internet
• Learn how domain names are used and managed on the
Internet
• Learn how servers and Internet security appliances can
be used to enable and improve the performance and
reliability of a network and the Internet
• Compare bandwidth technologies used on LANs, WAN,
and the Internet
 Physically and Logically
Dividing a Large Network
• You can divide a network using two
approaches:
1. You can physically divide the network using
hardware devices
2. You can logically divide the network using
software settings
• The first approach is called segmenting
the network; the second approach is
called subnetting.
 Bridges and Switches
• Bridges and switches are more intelligent than
hubs and make decisions involving whether to
allow traffic to pass or where to route that traffic,
reducing traffic on each segment and improving
network performance.
• A routing table is a database stored within a
router that is used to find the best network path
on which to forward information.
• A network bridge keeps routing tables for each
network to which it connects.
 Bridges and Switches
(Continued)
• The tables start out empty and all data packets
that reach the bridge from one segment are
passed on to the other segment connected to the
bridge.
• Just like bridges, switches keep tables of the
MAC addresses of all the devices connected to
the switch.
• Switches use these tables to determine which
path to use when sending packets.
• However, unlike a bridge, a switch passes a
packet only to its destination segment instead of
to all segments other than the one it came from.
 Bridges and Switches
(Continued)
• Bridges and switches use MAC addresses to
subdivide a network into physical segments.
• However, all the segments are still logically a
single network because each host is
communicating with other hosts on other
segments using the MAC address rather than
the IP address.
• As far as a host is concerned, it is not aware that
a bridge or a switch exists in the network.
 Subnetting
• A large network can be logically divided
into two or more networks based on IP
addresses rather than MAC addresses to
reduce congestion.

• Each division is called a subnet and the

process is called subnetting.
 Subnet Masks
• How does the host know if a remote host is on
the same network?
• An IP address is made up of the network ID and
the host ID.
• The host is told what portion of the IP address
identifies the network by an entry in the TCP/IP
configuration settings.
• This entry is called the network mask, or subnet
mask, and is used to define which portion of an
IP address identifies the network and which
portion identifies the host.
 Subnet Masks (Continued)
• The network mask is a group of four 8-bit
numbers separated by periods.
 Subnet Masks (Continued)
• If the network IDs had been different, the
host would not have attempted to resolve
the IP address to the MAC address, but
would have sent the data to the gateway
to its network.
• A gateway is any device, typically a router,
that provides access to another network.
• Subnet masks usually are not displayed as
32 bits separated by periods.
 Selecting a Subnet Mask
• A network engineer carefully selects a subnet
mask based on the number of subnets he needs
and the number of hosts planned for each
subnet.
• Table 6-3 on page 351 shows several examples
of subnet masks and explains the number of
hosts and subnets that can use each subnet
mask.
• Subnetting is necessary when a large company
is using a Class A, B, or C license for its entire
network and wants to use that one license over
several networks to prevent network congestion.
How Data Travels Across
Interconnected Networks
 How Data Travels Across
Interconnected Networks
• Networks are connected by routers, which
belong to more than one network.

• A router is a stateless device, meaning

that it is unconcerned about the data that it
is routing, but it is concerned about the
destination address of that data.
 Routers
• Routers are responsible for helping data
travel across interconnected networks.
• A router can forward data to the correct
network in a way that is similar to a
switch’s method.
• A router uses the most efficient path
available to forward packets to their
destination, which may be located across
a great geographical distance.
 Routers (Continued)
• A router can transmit a data packet to a
remote network only if the higher-layer
protocol that was used to produce the data
packet can be routed to a remote network.
• A brouter, short for bridge and router,
functions as both a bridge and a router.
• The device can forward routable protocols,
including TCP/IP and IPX/SPX packets,
and in these cases, is working as a router.
 Routers (Continued)
• Packets that are not routable, such as NetBEUI
packets, are forwarded to other local networks in
the manner that a bridge would forward packets.
• Like switches, routers use tables to determine
the best route by which to send the data to its
destination.
• When routers communicate with other routers to
build routing tables and determine availability of
routes, one of several protocols is used: RIP,
OSPF, BGP, DVMRP, NLSP, or IGRP.
 TCP/IP Routing
• Suppose a host computer wants to send
data to another host.
• Remember that the host uses its subnet
mask to decide if the destination host is on
its own or another network.
• If the first host knows that the remote host
is on its same network, it must discover
the MAC address of the remote host.
 TCP/IP Routing (Continued)
• If the sending host determines that the
remote host is on a different network, it
sends the data to the router, which is
serving as the gateway to remote
networks.
• When a packet arrives at a router, the
router decides if the packet belongs to a
host within its own local network or needs
to be routed to a different network.
Routing Across Many Networks
• For routing across interconnected
networks, each time a packet encounters
a router, its TTL is reduced by one.

• If the router must send the packet over a

network that cannot handle large packets,
the router divides the packet into smaller
packets.
 Default Gateways
• Sometimes, a large network has more
than one router, and so the network has
more than one gateway to other networks.
• Referring to textbook-figure 6.2: The
network in the upper-left is 250.1.2 and
has two routers (D and E), each of which
also belongs to other networks.
 Default Gateways (Continued)
• Host E is designated as the default
gateway, meaning that hosts on the
250.1.2 network send packets addressed
to other networks to this gateway first.
• The other router Host D, is called the
alternate gateway and is used if
communication to the default gateway
fails.
 Domain Names on the Internet
• Domain names are assigned because IP
address numbers are difficult to remember and
because companies might want to change their
IP addresses without also changing the Internet
name by which the outside world knows them.
• The last segment, or suffix, of a domain name is
called the top-level domain and tells you
something about the function of the host.
• The first word in a domain name is used to
identify a subcategory within the domain and is
called a canonical name, or CNAME.
 Assigning and Tracking Domain
Names and IP Addresses
• The organization responsible for overseeing this
operation is the IANA (Internet Assigned
Numbers Authority).
• Beginning in the spring of 1999, the responsibility
for assigning and tracking domain names and IP
addresses was transitioned from IANA to a
nonprofit, private sector organization regulated by
the U.S. Department of Commerce called ICANN
(Internet Corporation for Assigned Names and
Numbers).
• A company that can register these names and
numbers must be approved by ICANN and is
called a registrar.
 Domain Name Resolution
• Domain names and IP addresses do not have to
be permanently related.
• Two name resolution services track relationships
between domain names and IP addresses: DNS
(Domain Name System, also called Domain
Name Service) and Microsoft WINS (Windows
Internet Naming Service).
• DNS is the more popular of the two because it
works on all platforms.
• At the heart of DNS is a distributed database,
which initially must be created manually.
 How DNS Works
• DNS has three logical components:
– Computers searching for the IP address for a
domain name, called resolvers
– Servers that contain the information relating
domain names to IP addresses, called name
servers
– The databases of information needed to
resolve domain names and IP addresses,
called namespaces
 How DNS Works (Continued)
• The process of discovering an IP address for a
given domain name is called address
resolution.
• It is also possible to find the domain name for a
given IP address; this process is called reverse
resolution, or reverse mapping.
• Name servers are organized from the top down.
• Network Solutions maintains servers called root
servers that act as the highest level of authority
when locating domain name information.
 How DNS Works (Continued)
• A network that supports DNS has two or
more name servers, called the primary
name server and secondary name
server.
• The secondary server gets its information
from the primary server, and is sometimes
called the slave name server.
• An authoritative name server is the
server that has the most current
information about a domain name.
 How DNS Works (Continued)
• The group of networks for which the name
server is responsible collectively is called
the name server’s zone.
• A zone also can have a caching-only
server that does not keep authoritative
information, but only caches information
as it is used in case it is needed again
within a short period of time.
 DNS Records
• Each name server holds a piece of the
namespace, which is the database
containing information needed to resolve
domain names and IP addresses.
• A name server keeps the entries for each
domain name that it knows about in a
resource record, or DNS record.
 Directory Server
• A directory server stores information
about people, hosts, and other resources
on the network in directories and provides
this information to computers on the
network.
• The information in a directory is read more
often than it is written.
 How Directories Work
• Directories follow an upside-down tree
structure with the root at the top and
branches underneath the root in a
hierarchical structure.
• Directory servers sometimes use a protocol
called LDAP (Lightweight Directory
Access Protocol) to access directories.
• LDAP was designed to run over TCP and
can be used on the Internet or on an
intranet.
 How Directories Work
(Continued)
• LDAP is a “lighter” version of DAP (Directory
Access Protocol); LDAP has less code than
DAP.
• Another important directory standard is X.500,
which specifies how global directories can be
structured.
• X.500 directories are designed to provide a
listing of people within an organization so that
anyone with Internet access can look someone
up by country, organization, organizational level,
or name.
 Using Directories
• Directories and directory servers can
serve various functions on networks
and on the Internet.
• Directories on the Web are similar to
search engines in the way they
operate and provide information.
 Cache Servers
• Microsoft Internet Explorer supports
browser caching, which allows the user
to indicate how much hard drive space
should be allocated to Web caching.

• A cache server improves performance by

caching data so that the number of
requests to the Internet is reduced.
 Cache Servers (Continued)
• Cache servers save Web pages and other
files that users have requested so that
when a page is requested again, it can be
retrieved without accessing the Internet.
• Cache servers are placed between users
and the Internet.
• A cache server can run on a system such
as a proxy server or a router, or it can be
set up as a dedicated computer system.
 Mirrored Servers
• A mirrored server carries the same data
and services as another server. These
servers are exact replicas of the main
servers that they mirror and are updated
often to ensure that they contain the same
data.
• Mirrored servers have two main purposes:
– They reduce download time for users by
handling some of the traffic for a frequently
accessed Web site.
– They serve as backups for the main server in
case it goes down.
 Using a Mirrored Server
to Handle Site Traffic
• Web sites that get a lot of traffic often need
more than one server.
• If the traffic comes from different parts of
the world, it might be necessary to have
servers in different locations to provide the
best service for international customers.
 Using a Mirrored Server
as a Backup Server
• A mirrored server acts as a very effective
backup system.
• If the main server goes down, it is faster
and easier to switch operation to a
mirrored server than it is to restore
information from disks and tapes.
• When a mirrored server is used as a
backup for a server, it constantly copies
short segments of files from the main
server as they are updated.
 Print Servers
• Print servers make printers available for
shared use across a network or even
across the Internet.
• Each print server can have several
printers attached to it, and you can have
more than one print server on a network,
depending on the size of the network and
the needs and locations of the users on it.
 LPD Servers
• LPD (Line Printer Daemon) server is
print server software that initially was
developed on UNIX servers to handle print
jobs, but is now supported by Windows
Server 2003, Windows 2000 Professional,
and Windows NT Server.
• A client communicating with an LPD server
can use two protocols, LPR and LPQ,
which are part of the TCP/IP protocol
suite.
 LPD Servers (Continued)
• The client uses the LPR (Line Printer
Remote) protocol to send print jobs over a
TCP/IP network to the server.
• Clients that use LPR are sometimes
referred to as LPR clients.
• A second protocol, LPQ (Line Printer
Queue), is required for users to be able to
check on the status of print jobs they have
sent.
 IPP
• Another useful and more recently developed
printing protocol is IPP (Internet Printing
Protocol).
• IPP, which also enables printing across LANs and
the Internet, is generally more versatile than LPD
and its associated services, and is more easily
compatible with various operating systems.
• The greatest benefit of IPP is being able to find a
printer by using the printer’s IP address or URL.
• With IPP, you can find any Internet-connected
printer, print to it, and check the status of your
print job.
 Internet Security Appliances
• Internet security appliances, once called an
Internet-in-a-box, are becoming a popular
Internet access solution for small businesses.
• These devices combine a variety of
technologies, such as Internet prevention, and
other networking capabilities, into a single easy-
to-manage unit.
• Remote Access Service (RAS) provides a way
for a remote user to log on to the network using
telephone lines and a modem.
 Bandwidth Technologies
• Much attention is given to the amount of
data that can travel over a given
communication system in a given amount
of time.
• This measure of data capacity is called
bandwidth, also called data throughput
or line speed.
• The greater the bandwidth, the faster the
communication.
 Bandwidth Technologies Used
to Connect to an ISP
• A local area network has much less need
for data throughput than does a national
backbone.
• Lying between these two extremes on the
spectrum are many types of systems that
require varying degrees of bandwidth.
 Regular Telephone Lines
• Regular telephone lines, the most
common way to connect to an ISP, require
an internal or external modem.
• When data packets travel over telephone
lines, the Data Link layer protocol used is
PPP or SLIP.
• PPP (Point-to-Point Protocol) most often
is used to transmit TCP/IP packets from a
computer connected to an ISP or intranet
access point by telephone line.
 Cable Modem
• Cable modem communication uses cable
lines that already exist in millions of
households in the United States.
• Just as with cable TV, cable modems are
always connected.
• A cable modem is an example of
broadband media.
• Broadband refers to any type of
networking media that carries more than
one type of transmission.
PPPoE (Point-to-Point Protocol
over Ethernet)
• PPoE (Point-to-Point Protocol over Ethernet) is
a protocol that adapts PPP to work with Ethernet.
• PPPoE describes how the computer is to interact
with the converter box or modem when the two are
connected by an Ethernet cable connected to an
Ethernet network card in the computer.
• PPPoE gives the user the security and
authentication that is offered with PPP.
• PPPoE also sets standards for networks to
connect to the Internet via DSL modems and other
high-speed access services.
 ISDN
• ISDN (Integrated Services Digital
Network) is a technology developed in the
1980s that uses regular telephone lines,
and is accessed by a dial-up connection.

• ISDN is actually an early implementation

of DSL.
 DSL
• In the race to produce a fast data transmission
technology that is affordable for home use and
that offers a direct connection rather than a dial-
up connection, the telephone industry has
developed several similar technologies that
collectively are called DSL (Digital Subscriber
Line).
• Table 6-7 on page 390 lists common variations
of DSL.
• The most popular version of DSL is ADSL
(Asymmetric Digital Subscriber Line), which
is 50 times faster than ISDN and is direct
connect.
 Satellite Connections to the
Internet
• People who live in remote areas and want high-
speed Internet connections often are limited in their
choices.
• DSL and cable modems might not work where they
live, but satellite access is available from almost
anywhere.
• Technology is even being developed to use
satellites to offer Internet access on commercial
airlines.
• New technology allows data to be transmitted both
ways over the satellite so that telephone line
connections are not needed.
 Wireless Connections
• Wireless is an important technology for mobile
devices and for Internet access in remote
locations where other methods are not an
option.
• For Internet access, two popular applications of
wireless are fixed-point wireless, sometimes
called Wireless Local Loop (WLL), and mobile
wireless.
• With fixed-point wireless, an antenna sits on
your house or office building and communicates
with a base station antenna.
 Using Wireless Technology
to Browse the Internet
• Most of the wireless devices that advertise Internet
access are menu-driven, which means that you
select where you want to go from a menu instead
of typing in a URL.
• After the browser has been launched, you might
be charged for the time you are connected.
• Most Web sites still are not designed with wireless
technology in mind, and some devices run
software that converts HTML so it can be
displayed on wireless devices.
• This conversion process is called clipping
because it takes out, or clips, the images and
leaves a site with all text that uses menus to
navigate.
 Wireless Application Protocol
• The high demand for wireless access to the
Internet, including Web pages and e-mail, led to
the creation of the WAP (Wireless Application
Protocol).
• One goal of WAP is to bridge the gap between
the needs of traditional Internet access devices
and wireless Internet access devices.
• WAP is a communication standard designed for
mobile Internet access.
• One feature of WAP is WML (Wireless Markup
Language).
 Wireless Application Protocol
(Continued)
• WML is a markup language that is very
similar to HTML, but it is derived from XML
(Extensible Markup Language).
• WML files are called decks and are
divided into cards.
• Cards are sections of the deck that fit
onto one screen.
 Uses of Wireless Devices
• One popular service that is being provided
to customers with wireless Internet
devices is instant notification, or alerts.

• If you were invested heavily in a particular

stock, for example, wouldn’t it be nice to
be notified immediately if the value
changed?
 T Lines and E Lines
• The first successful system that supported
digitized voiced transmission was introduced in
the 1960s and was called a T-carrier.
• A T-carrier works with a leased digital
communications line provided through a
common carrier, such as Bellsouth or AT&T.
• The leased lines are permanent connections
that use multiplexing, a process of dividing a
single channel into multiple channels that can be
used to carry voice, data, video, or other signals.
T Lines and E Lines (Continued)
• The E-carrier is the European equivalent
of the American T-carrier.
• The E-carrier is a digital transmission
format devised by ITU at www.itu.int.
• A fractional T1 line is an option for
organizations that don’t need a full T1 line.
• The fractional T1 allows businesses to
lease some of the channels of a T1 line
rather than leasing all 24 channels.
 X.25 and Frame Relay
• Both X.25 and frame relay are packet-switching
communication protocols designed for long-
distance data transmission rather than the circuit-
switching technology used by the telephone
system.
• Both X.25 and frame relay use a PVC (permanent
virtual circuit).
• PVC is a permanent, logical connection between
two nodes.
• PVCs are not dedicated lines, like the T-carriers.
• The biggest advantage of X.25 and frame relay is
that you only have to pay for the amount of
bandwidth you require.
 ATM
• ATM (Asynchronous Transfer Mode) is a very fast
network technology that can be used with LANs, as well
as WANs.
• It uses fixed-length packets, called cells, to transmit
data, voice, video, and frame relay traffic.
• ATMs also use virtual circuits, meaning that the two
endpoints are stationary, but the paths between these
two endpoints can change.
• They can use either PVCs or SVCs.
• SVCs (switched virtual circuits) are logical, point-to-
point connections that depend on the ATM to decide the
best path to send the data.
 Mesh Topology
• A mesh topology provides multiple point-
to-point links between routers in a wide
area network, giving more than one choice
on how data can travel from router to
router.
• In a mesh topology, a router searches out
multiple paths and determines the best
path to take.
 Summary
• To relieve congestion, a network can be
segmented into smaller networks by using
a bridge, switch, or router.
• A large network logically can be divided
into subnets by using a subnet mask,
which takes a few bits from the network
portion of the IP addresses on a network
to define the subnets on the network.
 Summary (Continued)
• Domain names are an easy way to
remember an IP address, but also can be
assigned to different folders on a host so
that a host can have many domain names
assigned to it.
• Web caching can be used to store
frequently used Web pages in a temporary
place to decrease download time.
• Packet-switching divides data into packets
and sends each packet independently.